Normative Data and Conversion Equation for Spectral-Domain Optical Coherence Tomography in an International Healthy Control Cohort

Spectral-domain (SD-) optical coherence tomography (OCT) can reliably measure axonal (peripapillary retinal nerve fiber layer [pRNFL]) and neuronal (macular ganglion cell + inner plexiform layer [GCIPL]) thinning in the retina. Measurements from 2 commonly used SD-OCT devices are often pooled together in multiple sclerosis (MS) studies and clinical trials despite software and segmentation algorithm differences; however, individual pRNFL and GCIPL thickness measurements are not interchangeable between devices. In some circumstances, such as in the absence of a consistent OCT segmentation algorithm across platforms, a conversion equation to transform measurements between devices may be useful to facilitate pooling of data. The availability of normative data for SD-OCT measurements is limited by the lack of a large representative world-wide sample across various ages and ethnicities. Larger international studies that evaluate the effects of age, sex, and race/ethnicity on SD-OCT measurements in healthy control participants are needed to provide normative values that reflect these demographic subgroups to provide comparisons to MS retinal degeneration.

Original Contribution Section Editors: Clare Fraser, MD Susan Mollan, MD Normative Data and Conversion Equation for SpectralDomain Optical Coherence Tomography in an International Healthy Control Cohort Rachel Kenney, PhD, Mengling Liu, PhD, Lisena Hasanaj, MPA, Binu Joseph, MBBS, Abdullah A. Al-Hassan, MD, Lisanne Balk, PhD, Raed Behbehani, MD, Alexander U. Brandt, MD, Peter A. Calabresi, MD, Elliot M. Frohman, MD, PhD, Teresa Frohman, PA-C, Joachim Havla, MD, Bernhard Hemmer, MD, Hong Jiang, MD, Benjamin Knier, MD, Thomas Korn, MD, Letizia Leocani, MD, Elena H. Martínez-Lapiscina, MD, PhD, Athina Papadopoulou, MD, Friedemann Paul, MD, Axel Petzold, MD, PhD, Marco Pisa, MD, Pablo Villoslada, MD, Hanna Zimmermann, M.Eng, Hiroshi Ishikawa, MD, Joel S. Schuman, MD, Gadi Wollstein, MD, Yu Chen, MPH, PhD, Shiv Saidha, MBBCh, MD, MRCPI, Lorna E. Thorpe, MPH, PhD, Steven L. Galetta, MD, Laura J. Balcer, MD, MSCE, on behalf of the IMSVISUAL Consortium Background: Spectral-domain (SD-) optical coherence tomography (OCT) can reliably measure axonal (peripapillary retinal nerve fiber layer [pRNFL]) and neuronal (macular gan- glion cell + inner plexiform layer [GCIPL]) thinning in the retina. Measurements from 2 commonly used SD-OCT devices are often pooled together in multiple sclerosis (MS) studies and Departments of Neurology (RK, LH, BJ, SLG, LJB) and Population Health (RK, ML, YC, LET, LJB), New York University Grossman School of Medicine, New York, New York; Al-Bahar Ophthalmology Center (AAA-H, RB), Ibn Sina Hospital, Kuwait City, Kuwait; Centre for Research on Sports in Society (LB), Mulier Institute, Utrecht, Netherlands; Experimental and Clinical Research Center (AUB, AP, FP, HZ), Max Delbrueck Center for Molecular Medicine and Charité—Universitätsmedizin Berlin, Berlin, Germany; Department of Neurology (AUB), University of California, Irvine, California; Department of Neurology (PAC, SS), Johns Hopkins University, Baltimore, Maryland; Laboratory of Neuroimmunology (EMF, TF), Stanford University School of Medicine, Palo Alto, California; Institute of Clinical Neuroimmunology (JH), LMU Hospital, Ludwig Maximilians Universität München, Munich, Germany; Data Integration for Future Medicine consortium (DIFUTURE) (JH), Ludwig-Maximilians University, Munich, Germany; Department of Neurology (BH, BK, TK), Klinikum rechts der Isar, School of Medicine, Technical University of Munich, Munich, Germany; Munich Cluster for Systems Neurology (SyNergy) (BH, TK), Munich, Germany; Department of Neurology (HJ), Bascom Palmer Eye Institute, University of Miami Miller School of Medicine, Miami, Florida; Vita-Salute University & Hospital San Raffaele (LL, MP), Milano, Italy; Center of Neuroimmunology and Department of Neurology (EHM-L, PV), Hospital Clinic of Barcelona, Institut d’Investigacions Biomèdiques August Pi Sunyer (IDIBAPS), University of Barcelona, Barcelona, Spain; Neurologic Clinic and Policlinic (AP), MS Center and Research Center for Clinical Neuroimmunology and Neuroscience (RCN2NB) Basel, University Hospital Basel and University of Basel, Basel, Switzerland; NeuroCure Clinical Research Center (FP, HZ), Charité—Universitätsmedizin Berlin, Berlin, Germany; Moorfields Eye Hospital (AP), London, United Kingdom; The National Hospital for Neurology and Neurosurgery (AP), Queen Square, UCL Institute of Neurology, London, United Kingdom; Dutch Neuro-Ophthalmology Expertise Centre (AP), Amsterdam UMC, Amsterdam, the Netherlands; Oregon Health and Science University (HI), Portland, Oregon; Department of Ophthalmology (JSS, GW, SLG, LJB), New York University Grossman School of Medicine, New York, New York; Departments of Biomedical Engineering and Electrical and Computer Engineering (JSS), Tandon School of Engineering, New York University, Brooklyn, New York; Center for Neural Science (JSS), NYU, New York, New York; and Neuroscience Institute (JSS), NYU Langone Health, New York, New York. Supported in part by the NYU School of Medicine. P. A. Calabresi is supported by R01NS082347. J. Havla is (partially) funded by the German Federal Ministry of Education and Research (Grant Numbers 01ZZ1603[A-D] and 01ZZ1804[A-H] [DIFUTURE]). B. Hemmer received funding for the study by the European Union’s Horizon 2020 Research and Innovation Program (grant MultipleMS, EU RIA 733161) and the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy within the framework of the Munich Cluster for Systems Neurology (EXC 2145 SyNergy—ID 390857198). B. Knier is funded by the Else Kröner-Fresenius-Stiftung (Else KrönerFresenius Exzellenzstipendium). T. Korn is funded by the DFG (SFB1054-B06, TRR128-A07, TRR128-A12, TRR274-A01, Synergy Cluster of Excellence, EXC 2145, ID 390857198) and by the ERC (CoG 647215). J. S. Schuman received funding from the National Institutes of Health (Bethesda, MD) R01-EY013178. An unrestricted grant from Research to Prevent Blindness (New York, NY) to the Department of Ophthalmology, NYU Langone Health, NYU Grossman School of Medicine, New York, NY. P. Villoslada is funded by Instituto de Salud Carlos III, Spain and Fondo Europeo de Desarrollo Regional (PI15/0061). Address correspondence to Laura J. Balcer, MD, MSCE, Department of Neurology, NYU School of Medicine, 222 E. 41st Street, New York, NY 10017; E-mail: [e-mail] 442 Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution clinical trials despite software and segmentation algorithm differences; however, individual pRNFL and GCIPL thickness measurements are not interchangeable between devices. In some circumstances, such as in the absence of a consistent OCT segmentation algorithm across platforms, a conversion equation to transform measurements between devices may be useful to facilitate pooling of data. The availability of normative data for SD-OCT measurements is limited by the lack of a large representative world-wide sample across various ages and ethnicities. Larger international studies that evaluate the effects of age, sex, and race/ethnicity on SD-OCT measurements in healthy control participants are needed to provide normative values that reflect these demographic subgroups to provide comparisons to MS retinal degeneration. Methods: Participants were part of an 11-site collaboration within the International Multiple Sclerosis Visual System (IMSVISUAL) consortium. SD-OCT was performed by a trained technician for healthy control subjects using Spectralis or Cirrus SD-OCT devices. Peripapillary pRNFL and GCIPL thicknesses were measured on one or both devices. Automated segmentation protocols, in conjunction with manual inspection and correction of lines delineating retinal layers, were used. A conversion equation was developed using structural equation modeling, accounting for clustering, with healthy control data from one site where participants were scanned on both devices on the same day. Normative values were evaluated, with the entire cohort, for pRNFL and GCIPL thicknesses for each decade of age, by sex, and across racial groups using generalized estimating equation (GEE) models, accounting for clustering and adjusting for within-patient, intereye correlations. Change-point analyses were performed to determine at what age pRNFL and GCIPL thicknesses exhibit accelerated rates of decline. Results: The healthy control cohort (n = 546) was 54% male and had a wide distribution of ages, ranging from 18 to 87 years, with a mean (SD) age of 39.3 (14.6) years. Based on 346 control participants at a single site, the conversion equation for pRNFL was Cirrus = 25.0 + (1.0 · Spectralis global value). Based on 228 controls, the equation for GCIPL was Cirrus = 24.5 + (0.9 · Spectralis global value). Standard error was 0.02 for both equations. After the age of 40 years, there was a decline of 22.4 mm per decade in pRNFL thickness (P , 0.001, GEE models adjusting for sex, race, and country) and 21.4 mm per decade in GCIPL thickness (P , 0.001). There was a small difference in pRNFL thickness based on sex, with female participants having slightly higher thickness (2.6 mm, P = 0.003). There was no association between GCIPL thickness and sex. Likewise, there was no association between race/ethnicity and pRNFL or GCIPL thicknesses. Conclusions: A conversion factor may be required when using data that are derived between different SD-OCT platforms in clinical trials and observational studies; this is particularly true for smaller cross-sectional studies or when a consistent segmentation algorithm is not available. The above conversion equations can be used when pooling data from Spectralis and Cirrus SD-OCT devices for pRNFL and GCIPL thicknesses. A faster decline in retinal thickness may occur after the age of 40 years, even in the absence of significant differences across racial groups. Journal of Neuro-Ophthalmology 2022;42:442–453 doi: 10.1097/WNO.0000000000001717 © 2022 by North American Neuro-Ophthalmology Society L. J. Balcer is editor-in-chief of the Journal of Neuro-Ophthalmology. A. U. Brandt is named as inventor on several patents and patent applications describing multiple sclerosis serum biomarkers, retinal image analysis methods, and human pose estimation methods. He is cofounder and holds shares in companies Motognosis and Nocturne. P. A. Calabresi is PI on grants to Johns Hopkins from Principia and Genentech. He has consulted for Disarm, Nerveda, Biogen, and Avidea. E. M. Frohman has received consulting and speaker fees from Novartis, Genzyme, Biogen, Alexion, and Janssen. T. Frohman has received consulting fees from Alexion. S. L. Galetta has been a consultant for Biogen and Genentech. J. Havla reports grants for OCT research from the Friedrich-Baur-Stiftung and Merck, personal fees and nonfinancial support from Celgene, Merck, Alexion, Novartis, Roche, Santhera, Biogen, Heidelberg Engineering, and Sanofi Genzyme, and nonfinancial support of the Guthy-Jackson Charitable Foundation, all outside the submitted work. B. Hemmer has served on scientific advisory boards for Novartis; he has served as DMSC member for AllergyCare, Polpharma, and TG therapeutics; he or his institution have received speaker honoraria from Desitin; his institution received research grants from Regeneron for MS research. He holds part of 2 patents: 1 for the detection of antibodies against KIR4.1 in a subpopulation of patients with MS and 1 for genetic determinants of neutralizing antibodies to interferon. All conflicts are not relevant to the topic of the study. B. Knier received travel support and a research grant from Novartis (Oppenheim research award). F. Paul serves as an Associate Editor for Neurology: Neuroimmunology & Neuroinflammation, reports research grants and speaker honoraria from Bayer, Teva, Genzyme, Merck, Novartis, and MedImmune and is member of the steering committee of the OCTIMS study (Novartis). A. Papadopoulou has received speaker fee from Sanofi-Genzyme and travel support from Bayer AG, Teva, and Hoffmann-La Roche. Her research was/is being supported by the University and University Hospital of Basel, the Swiss Multiple Sclerosis Society, the “Stiftung zur Förderung der gastroenterologischen und allgemeinen klinischen Forschung sowie der medizinischen Bildauswertung,” and the Swiss National Science Foundation (Project number: P300PB_174480). A. Petzold is supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology. A. Papadopoulou is part of the steering committee of the ANGI network that is sponsored by ZEISS and steering committee of the OCTiMS study that is sponsored by Novartis and reports speaker fees from Heidelberg-Engineering. S. Saidha has received consulting fees from Medical Logix for the development of CME programs in neurology and has served on scientific advisory boards for Biogen, Genzyme, Genentech Corporation, EMD Serono & Celgene. He has consulted for Carl Zeiss Meditec. He is the PI of investigator-initiated studies funded by Genentech Corporation and Biogen Idec, and received support from the Race to Erase MS foundation. He has received equity compensation for consulting from JuneBrain LLC, a retinal imaging device developer. J. S. Schuman: Aerie Pharmaceuticals, Inc.: consultant/advisor, equity owner. BrightFocus Foundation: grant support. Boehringer Ingelheim: consultant/advisor. Carl Zeiss Meditec: patents/royalty/consultant/advisor. Massachusetts Eye and Ear Infirmary and Massachusetts Institute of Technology: intellectual property. National Eye Institute: grant support. New York University: intellectual property. Ocugenix: equity owner, patents/royalty. Ocular Therapeutix, Inc.: consultant/advisor, equity owner. Opticient: consultant/advisor, equity owner. Perfuse, Inc.: consultant/advisor. Regeneron, Inc.: consultant/advisor. SLACK Incorporated: consultant/advisor. Tufts University: intellectual property. University of Pittsburgh: intellectual property. P. Villoslada holds stocks and has received consultancy fees from Accure therapeutics, Spiral Therapeutics, QMENTA, Attune Neurosciences, CLight, NeuroPrex, and Adhera Health. H. G. Zimmermann received research grants from Novartis and speaking honoraria from Bayer Healthcare. The remaining authors report no disclosures. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the full text and PDF versions of this article on the journal’s Web site (www.jneuro-ophthalmology.com). Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 443 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution I n neurology, optical coherence tomography (OCT) has become an important method to determine integrity of visual pathway. One example is multiple sclerosis (MS), with OCT studies for over 10 years. As more studies have been published describing the association of spectral-domain optical coherence tomography (SD-OCT) measurements with brain atrophy and with reductions in visual function test scores, the use of OCT as a structural outcome measure in clinical trials of MS therapies is more widely adopted (1). With collaborations of neuro-ophthalmologists and MS specialists, the use of multiple OCT devices has created challenges for synergizing sites for trials. Often, clinical trials will use one or both of the 2 most common SD-OCT platforms (Cirrus OCT or Spectralis); however, individual peripapillary retinal nerve fiber layer (pRNFL) and ganglion cell + inner plexiform layer (GCIPL) thickness measurements are not interchangeable between these devices (2– 5) unless consistent segmentation algorithms are used or the sample size is large enough to overcome systematic differences. SD-OCT measurements have high levels of reproducibility and low degrees of variability within each of the SD-OCT platforms for healthy control participants (6) and in MS participants (7,8) across multicenter studies (9). Measurements from the Cirrus and Spectralis SD-OCT devices are often pooled together despite differences in software algorithms, hardware/optical components, and segmentation areas (3,10). Equations that convert pRNFL and GCIPL measurement between the 2 most widely used SD-OCT devices will be useful for clinical trials, clinical practice, and observational studies. Such an equation will provide a method to relate measurements obtained by Spectralis and Cirrus SDOCT platforms and to more accurately pool data. Previous investigations have estimated conversion equations for pRNFL thickness in healthy volunteers and in participants with glaucoma (3,11). The relation between GCIPL measurements for the Cirrus vs Spectralis OCT platforms has not been investigated. Normative data are important to provide a basis to compare retinal degeneration seen in MS. The availability of normative data for SD-OCT measurements is currently limited by the lack of a large world-wide representative sample of normative data across age groups. Current normative values are based on the following: 1) healthy control comparison groups from published studies that may be subject to selection bias and are not stratified by age, sex, or race; 2) normative values provided in the OCT device software by the manufacturers, whose analyses were based on small sample sizes (fewer than 300 subjects in the Cirrus OCT normative cohort), did not include macular scans (Spectralis normative cohort) and also do not account for the potential effects of sex, race, or ethnicity. A number of recent studies demonstrate meaningful variability in SD-OCT measurements by age, sex, and race, 444 suggesting that these current value models are potentially problematic (12,13). Thinner pRNFL measurements were associated with older age and Caucasian race in a study using one of the first versions of OCT technology (timedomain) (14). A study of people with MS by Kimbrough et al (ref) showed that African Americans (AAs) had higher baseline pRNFL thicknesses in the healthy control group compared with non-AA individuals. However, there were no differences in GCIPL values at baseline between AA participants and non-AA individuals in the healthy control group (15). This study had a small sample size (n = 14 AA healthy control participants). Another study with 31 AAs and 61 Caucasians found higher pRNFL values in the AA group, with no differences in GCIPL thickness between groups (16). Larger international studies that evaluate the effects of age, sex, and race/ethnicity on SD-OCT pRNFL and GCIPL measurements in healthy control participants are needed to provide normative values that represent various ages, sex, and demographic subgroups. The purpose of this investigation was to develop a conversion equation for pRNFL and GCIPL thicknesses to improve comparability of these measurements derived from the Cirrus OCT and Spectralis OCT devices in healthy control participants. We also sought to examine the effect of age, sex, and race/ethnicity on OCT measurements in a large international cohort of healthy control participants. METHODS Study Cohort OCT and high-contrast visual acuity measurements were collected for 546 healthy control participants. Participants were $18 years of age with no history of ocular or neurological disease; high-contrast visual acuities were better than 20/40, and refractive error was between 26 and +6 spherical diopters, inclusive. Participants were part of an 11-site collaboration within the IMSVISUAL (International MS Visual System) consortium in the United States, Europe, and the Middle East. IMSVISUAL is an international, collaborative group of researchers with over 140 members from 40 countries that investigate the visual pathway in MS and related demyelinating disorders (www.imsvisual.org) (17). Through concerted collaborative efforts, IMSVISUAL has facilitated high-quality, large-scale studies of the visual system in MS, including those examining the association of OCT measurements with future MS disability (18). IMSVISUAL has established guidelines for reporting OCT in published studies (19,20), and previous work from our group that established ideal intereye difference thresholds for using OCT measurements in the diagnosis of optic nerve lesions (21) have also been determined through a recent IMSVISUAL collaboration. For the present study, each site’s institutional review board approved the study procedures. All participants provided written informed Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution consent to participate in research studies at the individual sites. Data sharing agreements were completed between each study site and New York University (NYU) Grossman School of Medicine. Age at the time of visit was self-reported or calculated as the difference between date of birth and date of visit. Age was categorized in decades (,30, 30–39, 40–49, 50–59, 60–69, 70+ years). Race/ethnicity data were collected, and participants were self-categorized as non-Hispanic Caucasian, non-Hispanic African American, Hispanic, Asian or Pacific Islander, Alaskan or American Indian, or Other. These racial and ethnicity categories are based on the United States Census guidelines. However, because very few participants self-identified as Hispanic, separating Hispanic Caucasians and Hispanic African Americans was not feasible; therefore, Hispanic was defined as its own category. Optical Coherence Tomography Either Spectralis SD-OCT (Heidelberg Engineering, Heidelberg, Germany) or Cirrus OCT (Carl Zeiss Meditec, Dublin, CA), or both, were performed for all participants by a trained technologist as part of ongoing MS vision research studies at all sites. pRNFL thickness was measured using a 3.4-mm peripapillary ring scan on the Spectralis OCT. On the Cirrus OCT, optic nerve head (ONH) Cube 200 · 200 scans were used to measure pRNFL thickness in a 3.4-mm circle centered on the optic disc. Macular volume scans encompassing a .6-mm area surrounding the fovea were obtained using custom scans on the Spectralis OCT and by automated macular volume cube 200 · 200 or 512 · 128 (6 · 6 mm) scans on the Cirrus OCT. Macular GCIPL thickness was measured as the sum of the ganglion cell layer plus inner plexiform layer thicknesses and was collected for healthy control participants at 5 of the sites in the IMSVISUAL study. GCIPL measurements were obtained using the automated macular volume cube scans with a measurement area of a 4 · 5-mm annulus surrounding the fovea (7) on Cirrus OCT and from the macular volume scan encompassing a 6 · 6-mm cylinder surrounding the fovea on the Spectralis OCT. Automated segmentation protocols with manual inspection and correction of lines delineating retinal layers were used for all scans. Manual review of the OCT images by trained technicians and/or clinicians was performed to ensure that all scans met quality control standards. The OSCAR-IB criteria for scan quality control (22) were followed for all OCT scans. OCT results are reported in this article in concordance with the APOSTEL 2.0 guidelines (19,20). Cirrus and Spectralis OCT data were collected on the same day for the same participants at one study site (NYU). Repeated Cirrus OCT (2 or more scans with the same protocol on the same device) and repeated Spectralis OCT measurements were also collected in a subset of participants at the NYU site. Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 Statistical Analyses Conversion Equation Because 2 SD-OCT devices were used in these studies (devices chosen by study sites based on institutional preference and before the inception and design of these analyses), a conversion equation was developed using a structural equation model (Fig. 1). This model accounted for clustering because 2 eyes were included for each subject; the model was based on the 173 participants (346 eyes) for pRNFL and 114 participants (228 eyes) for GCIPL thicknesses who were scanned on the same day using both OCT devices and with repeated measures on one or both devices at one site (NYU). The maximum likelihood with missing values (MLMVs) method was used; this assumes joint normality of all variables and that missing values are missing at random and allows observations with missing values to be included in the model. Structural equation modeling (SEM), similar to a regression equation, provides the intercept and beta coefficient to convert between devices; however, the SEM also considers unmeasured error in the devices and is, therefore, more accurate than using regression modeling alone (23). Path (or regression) coefficients are a measure of association between the 2 devices and capture related FIG. 1. Structural equation model (SEM) model. The path diagram illustrates the SEM that describes the relationship of the repeated pRNFL measurements for each SD-OCT device with the unknown true global pRNFL thickness values. Cirrus is set as the reference value for comparison to Spectralis. pRNFL indicates peripapillary retinal nerve fiber layer; SD-OCT, spectral-domain optical coherence tomography. 445 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution changes between the 2 devices. The path coefficients were constrained to 1 and the intercepts were constrained to 0 for the Cirrus device, allowing Cirrus to be the reference device compared with Spectralis. The path coefficients for Spectralis measurements were constrained to be equal because they are inherent to the device and assumed not to vary between measures, as were the intercepts. The conversion equation was validated using leave-one-out cross validation (LOOCV) on an independent dataset of participants who had measurements on both devices on the same day but did not have repeated measures on either device and therefore were not included in the SEM developing the conversion equation. The equation was validated in both healthy control participants and people with MS separately. The independent dataset of healthy control participants included 95 eyes for pRNFL and 22 eyes for GCIPL. The independent dataset of people with MS included 37 eyes for pRNFL and 6 eyes for GCIPL (which may be too small to validate results for GCIPL in people with MS). Intraoperator reproducibility for each device was evaluated by the intraclass correlation coefficient (ICC) and coefficient of variation (CV). Once the conversion equation was developed, data were pooled together to increase sample size for further analysis because some sites used Cirrus OCT only and some sites used only Spectralis OCT measurements. Values from Cirrus device were used when those data were available, and values from Spectralis device were converted based on the conversion equation developed. This method was used to reduce systematic bias from having measurements from different devices with different image acquisition and segmentation algorithms; this also increases sample size and power of the study by allowing for pooling of data from the 2 OCT devices. ysis with one inflection point evaluates the differences in slope before and after the inflection point at multiple points to determine the point where the slopes have the greatest difference. Age was also dichotomized based on results of change-point analyses as a secondary analysis to look for differences in progression before and after the change point. Missing race data (11.5%) and GCIPL data (25.6%) were imputed using the MICE command in Stata 16.0 with 100 multiple imputations as a sensitivity analysis. Additional sensitivity analyses examining associations of SD-OCT measures with age, race/ethnicity, and sex were performed for data from each OCT device to ensure that trends seen were not affected by pooling of the data. Analyses were performed using Stata 16.0, R and Python software. Normative Data Analysis Intereye pRNFL and GCIPL differences were calculated by subtracting right eye values from left eye values and using the absolute value of the difference. Descriptive statistics report continuous variables as mean values and standard error (SE) and categorical variables as frequency and percentages. Mean values and standard errors were calculated using linear mixed effects model, which account for clustering because 2 eyes were used for each participant. The effects of age, sex, and race/ethnicity on pRNFL and GCIPL thickness were evaluated using generalized estimating equation (GEE) regression models accounting for within-subject, intereye correlations because both eyes of each participant were included in the model. Bland–Altman plots were created using Stata 16.0 to compare OCT thickness values between devices. Change-point analyses to determine inflection points where pRNFL and GCIPL thicknesses change with respect to their degrees decline were performed with the “segmented” package in R and confirmed in Python. Change-point anal- At the single site (NYU) used for development of the conversion equation and OCT reproducibility results, PRNFL measurements were performed on both Cirrus and Spectralis devices on 346 healthy control eyes and GCIPL measurements were performed on 228 healthy control eyes on both Cirrus and Spectralis devices. Number of eyes with repeated scans on each device are shown in Figure 2. Eyes with poor quality scans or ocular pathology were excluded. Both Cirrus and Spectralis platforms had excellent reproducibility for both pRNFL and GCIPL measurements (ICC ranging from 0.995 to 0.998), although the sample size for GCIPL on the Spectralis machine was small. Reproducibility results for this cohort are shown in Table 2. Bland–Altman plots showing agreement between devices for pRNFL and GCIPL thicknesses are shown in Figure 3. A conversion equation for pRNFL thickness between Cirrus and Spectralis SD-OCT was developed based on 346 healthy control eyes that were scanned on both OCT devices on the same day, with repeated measures on at least on device, at a single site (NYU) using structural equation 446 RESULTS Study Cohort Data from healthy control participants (n = 546) from 9 sites in the IMSVISUAL consortium were included in this study. Six healthy control participants in the cohort were excluded for not having age information available, resulting in a final pooled cohort size of 540 healthy control participants. The cohort had a fairly equal distribution of women and men (54% male), was primarily Caucasian (76%), and had a wide distribution of ages from ranging from 18 to 87 years, with a mean (SD) of 39.3 (14.6) years. Demographics of these participants are shown in Table 1. Distribution of racial groups by country is shown in Supplemental Digital Content (see Table 1, http://links.lww.com/WNO/A632). The non-US cohorts were comprised almost exclusively by Caucasians, reflecting the demographics of those countries. Conversion Equation Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 pRNFL (n = 546) Age, yr 18–29 30–39 40–49 50–59 60–69 70+ Sex Male Female Race Caucasian Hispanic African American Asian or Pacific Islander Other Missing Country Germany Holland Kuwait Spain Switzerland United States GCIPL (n = 406) Frequency (Number/Percent) Mean (SE)* Beta Coefficient (95% CI) P† 165 (30.2) 179 (32.8) 81 (14.8) 50 (9.2) 48 (8.8) 23 (4.2) 94.8 94.9 94.4 91.2 89.1 88.2 Reference 20.41 (22.51, 1.70) 20.65 (23.31, 2.00) 22.15 (25.35, 1.04) 26.11 (29.21, 23.01) 26.65 (210.68, 22.63) Reference 0.705 0.630 0.187 ,0.001 0.001 85.2 85.6 85.4 82.0 82.2 80.8 247 (45.2) 303 (54.9) 92.5 (0.6) 94.6 (0.5) Reference 2.60 (0.91, 4.29) Reference 0.003 413 (75.6) 11 (2.0) 22 (4.0) 29 (5.3) 8 (1.5) 63 (11.5) 93.6 97.6 94.0 92.3 90.6 94.2 (0.5) (3.1) (2.0) (1.5) (0.9) (1.1) Reference 3.89 (21.62, 9.41) 1.55 (22.49, 5.59) 21.18 (24.76, 2.41) 21.70 (28.16, 4.76) — 54 (9.9) 19 (3.5) 38 (7.0) 108 (19.8) 33 (6.0) 294 (53.9) 92.7 87.1 96.0 94.6 99.0 93.0 (1.2) (1.8) (1.0) (0.9) (1.6) (0.6) Reference 25.10 (210.06, 20.13) 2.50 (21.29, 6.28) 2.97 (20.01, 5.95) 5.53 (1.58, 9.49) 0.46 (22.37, 3.30) (0.7) (0.7) (1.0) (1.3) (1.7) (2.3) Mean (SE)* Beta Coefficient (95% CI) P† Reference 20.31 (21.68, 1.06) 20.51 (22.19, 1.17) 22.58 (24.71, 20.44) 24.81 (26.70, 22.92) 24.05 (26.44, 21.66) Reference 0.659 0.550 0.018 ,0.001 0.001 84.4 (0.4) 84.8 (0.4) Reference 20.32 (21.41, 0.77) Reference 0.562 Reference 0.166 0.453 0.520 0.607 — 85.3 82.5 81.8 83.2 79.9 83.1 Reference 20.79 (24.11, 2.53) 0.47 (22.09, 3.03) 0.83 (21.97, 3.63) 21.28 (24.49, 1.93) — Reference 0.640 0.718 0.560 0.434 — Reference 0.044 0.196 0.051 0.006 0.748 84.8 (0.5) — 83.7 (0.6) 88.6 (0.4) 93.3 (0.7) 81.9 (0.4) Reference — 22.05 (24.61, 0.50) 4.12 (1.94, 6.30) 7.48 (4.84, 10.12) 23.42 (25.58, 21.25) Reference — 0.115 ,0.001 ,0.001 0.002 (0.5) (0.5) (0.7) (1.4) (1.2) (1.5) (0.4) (2.2) (1.4) (1.6) (1.4) (0.8) *Mean values and standard errors calculated using linear mixed effects models accounting for clustering of 2 eyes per subject. † Beta estimates and P-values estimated using GEE models accounting for clustering of 2 eyes per subject. GCIPL, ganglion cell + inner plexiform layer; GEE, generalized estimating equation; pRNFL, peripapillary retinal nerve fiber layer; SE, standard error. Original Contribution 447 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. TABLE 1. Demographic and spectral-domain optical coherence tomography summary statistics and generalized estimating equation results for healthy control cohort Original Contribution FIG. 2. Flow chart of scans used in reproducibility and SEM analyses. The number of eyes (people) that were scanned for each individual measure on each device repeated twice on the same day is depicted in line (A). These scans were used to determine reproducibility results shown in Table 2. Line (B) shows the number of eyes (people) who were used in the SEM model. These people had at least one measure on both devices and repeated measures on at least one device. SEM indicates structural equation model. modeling. The conversion equation performed well when tested on an independent dataset for both healthy control participants (n = 95 eyes, R2 = 0.85, LOOCV) and people with MS (n = 37 eyes, R2 = 0.91, LOOCV). The conversion equation for pRNFL is as follows: ½Cirrus ¼ 2 5:0 þ ð1:0 · Spectralis global pRNFL valueÞ utively on each device on the same day. The equation performed well in an independent dataset of healthy control participants (n = 22 eyes, R 2 = 0.933, LOOCV) and people with MS (n = 6 eyes, R2 = 0.948, LOOCV), although the sample size may be too small to validate results. The conversion equation for GCIPL is as follows: ½Cirrus ¼ 2 4:5 þ ð0:9 · Spectralis global GCIPL valueÞ ½Cirrus ¼ 5:0 þ ð1:0 · Cirrus global pRNFL valueÞ The standard error for the equation was 0.02 and the overall SEM model had R2 = 0.994. The 95% confidence interval (CI) was 0.96–1.05 for the beta coefficient and 0.95–9.00 for the intercept. Similarly, for GCIPL, the conversion equation was developed on 228 healthy control eyes scanned consec- ½Spectralis ¼ 5:0 þ ð1:1 · Cirrus global GCIPL valueÞ The standard error for the equation was 0.02 and the overall SEM model had R2 = 0.996. The 95% CI was 0.83–0.93 for the beta coefficient and 28.35 to 20.58 for the intercept. TABLE 2. Reproducibility of optical coherence tomography measurements for healthy control cohort at single site (New York University) with repeated measures Number with repeated measures, eyes (people) Difference between repeated measures, microns, median (IQR/range) ICC (repeated measures) CoV (repeated measures) Difference between devices, microns, median (IQR/range) Cirrus pRNFL Spectralis pRNFL 206 (112) 1 (0–2) (0–8) 0.998 2.07% 310 (166) 1 (0–2) (0–6) 0.996 1.48% 5 (3–7) (0–17)* Cirrus GCIPL Spectralis GCIPL 203 (111) 33 (19) 0 0.4 (0–1) (0.2–0.7) (0–6) (0.0–2.7) 0.995 0.997 1.24% 1.16% 14 (12.5–15.3) (5.0–18.6)† *n = 439 eyes (224) people had at least one pRNFL measurements on both Spectralis and Cirrus devices. † n = 216 eyes (109) people had at least one GCIPL measurements on both Spectralis and Cirrus devices. CoV, coefficient of variation; GCIPL, ganglion cell + inner plexiform layer; ICC, intraclass correlation coefficient; IQR, interquartile range; pRNFL, peripapillary retinal nerve fiber layer. 448 Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 3. Bland–Altman plots. A. Bland–Altman plot for pRNFL thickness showing agreement between devices. The y axis shows the difference in measurement between Cirrus pRNFL and Spectralis pRNFL thickness. The x axis shows the pRNFL thickness. B. Bland–Altman plot for GCIPL thickness. GCIPL indicates ganglion cell + inner plexiform layer; pRNFL, peripapillary retinal nerve fiber layer. An example of using the equation and the predictive error of the 95% CI would be if we used a value of 100 mm for Spectralis GCIPL. Using the conversion equation, Cirrus would equal 85.5 mm. If we consider the lowest extreme of the 95% CI (28.35 intercept and 0.83 beta coefficient) and the highest (20.58 intercept and 0.93 beta coefficient), we would have Cirrus = 74.65 or 92.2 mm. This would be a difference of 10.8% or 6.7%, respectively, from the predicted value of 85.5 mm using the equation. Normative Data Data from the whole cohort (n = 540 healthy control participants) showed a mean pRNFL thickness of 93.8 (SD 9.9) microns and a mean GCIPL thickness of 84.6 (SD 6.7) microns. Mean pRNFL and GCIPL thicknesses were consistent over decades 18–29, 30–39, and 40–49 years and then showed a statistically significant decline for each subsequent decade (Table 1, Fig. 4). Similar results were seen for each device individually before pooling data together. Devicespecific normative values are shown in Supplemental Digital Content (see Tables 2 and 3, http://links.lww.com/WNO/ A633, http://links.lww.com/WNO/A634). Change-point analyses showed a transition point at age 40 years for pRNFL and 37 years for GCIPL thicknesses (Fig. 4). Age overall was associated with pRNFL decline at a rate of 21.31 mm per decade and a GCIPL decline at a rate of 1.05 mm per decade (P , 0.001, statistical test). When considering age as a dichotomous variable (above or below the age of 40 years) based on the results of change-point analysis, pRNFL decline was not associated with age below 40 years. However, there was an associated decline of 2.4 mm per decade above the age of 40 years (P , 0.001, GEE models adjusting for sex, race, and country). Similarly, GCIPL had a faster decline above the age of 40 years (1.4 mm per decade, P = 0.002, GEE models adjusting for sex, race, and country) and had no associated decline below the age of 40 years. Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 There were small differences in pRNFL thickness based on sex, with female participants having slightly higher thickness (by an average of 2.6 mm, P = 0.003, GEE models adjusting for age, race/ethnicity, and country). There was no association between GCIPL thickness and sex. Likewise, there were no associations between race/ethnicity and pRNFL or GCIPL thicknesses (Table 1). Sensitivity analyses imputing missing race (n = 63, 11.5%) and GCIPL thickness (n = 140, 25.6%) showed similar results. Likewise, analyses separating SD-OCT measures by device showed similar trends, suggesting that pooling of the data with the conversion equation did not potentially bias the results. DISCUSSION Results of this investigation demonstrate that a conversion factor is necessary when using 2 different SD-OCT devices clinically or in research studies. This study is unique because it pools together SD-OCT data on a large international multicenter study using a conversion factor. Furthermore, normative values for SD-OCT devices are presented in our investigations using a large, diverse, multicenter, international cohort. We observed a decline in SD-OCT-measured pRNFL thickness after age 40 years and saw slightly thicker pRNFL values in female participants. SD-OCT has emerged as an important tool for detecting optic neuropathies and in capturing axonal and neuronal degeneration in MS. It is important to be able to compare SD-OCT measurements across different OCT platforms both clinically and in research studies. It is also critical to understand normative SD-OCT values based on age and sex. The equation for conversion between Cirrus (C) and Spectralis (S) devices for pRNFL thickness (C = 25.05 + S) correlates well with the actual differences seen between the mean values of the groups (5 mm); its simplicity could make the equation highly useful for clinical practice and for 449 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 4. Boxplot for pRNFL thickness by age (A) and GCIPL thickness by age (B). Scatterplot with change-point analysis for pRNFL (C) and GCIPL (D). GCIPL indicates ganglion cell + inner plexiform layer; pRNFL, peripapillary retinal nerve fiber layer. observational research studies and clinical trials. The beta coefficient between the 2 devices was 1.0, suggesting the measurements are on the same scale, yet differing by 5 mm, with Spectralis being higher. Although an equation for GCIPL was also developed, the small sample size of participants who had GCIPL measurements repeated on the Spectralis device was much smaller (33 eyes) than for pRNFL and the results may be underpowered. These findings, coupled with the variability in measurement area (median 14 [IQR 12.5–15.3]) and segmentation algorithms, suggest that this equation will need to be tested further to determine if it is generalizable to other studies. Importantly, our findings emphasize the need for a correction factor when comparing OCT measurements across platforms. To overcome variability between devices for GCIPL measures, a consistent segmentation algorithm could be used on both devices in lieu of a correction factor. However, pRNFL measurement is fairly standardized; the conversion equation may be useful in clinical trials, clinical practice, and other observational research studies. In a previous study by Pierro et al, the pRNFL conversion factor between Cirrus and Spectralis was found to be Cirrus = 2.969 + 0.942 · Heidelberg Spectralis. This is similar to our equation when considering numbers close 450 to the mean for healthy control participants. Both equations show that Cirrus has a lower measurement than Spectralis at this scale. For example, a Spectralis measurement of 100 would equal a Cirrus measurement of 97.2 in Pierro’s study and 95.0 in our study. Our study had a much higher sample size (346 eyes vs 38 eyes), so our calculations may be more accurate. In this large, multicenter international study, there were differences by sex only for pRNFL thickness; these were not observed for GCIPL thickness. A large population-based study of 7,868 Caucasian healthy control participants in Germany also found the pRNFL to be thicker in female participants (with variations depending on scanning distance from the optic nerve and by quadrants), by 1 mm, whereas GCIPL was not evaluated (24). The Cirrus OCT Normative Database Study Group evaluated age and race differences for GCIPL on Cirrus OCT in 282 healthy control participants and found no difference between male and female participants after adjusting for axial length; they also found no differences by race after adjusting for age, axial length, and pRNFL thickness. These results are similar to those of the present study (25). There were no differences between by race or ethnicity for either pRNFL or GCIPL measurements in our study Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution cohort. Other reports have shown differences in retinal layer thicknesses between racial groups in both healthy control participants and in people with MS (15). A recent study of 31 African American (AA) and 61 Caucasian American (CA) healthy control participants showed that AAs had higher pRNFL thicknesses than CAs (P = 0.042). There were no differences in GCIPL thicknesses in that study (16). Another study found that Asians had greater thicknesses for average pRNFL (26). The Cirrus OCT Normative Database Study Group found thinner pRNFL values in Europeans compared with Hispanics and Asians (25). The sample sizes for racial/ethnic groups in our study may have been too small to detect a difference. The potential differences in pRNFL or GCIPL measurements will need to be examined further and in larger cohorts. Overall pRNFL and GCIPL thickness for the whole cohort are comparable with other studies (27,28). A decline in pRNFL and GCIPL thickness, with a faster rate of thinning beyond age 40 years, was seen in our present study cohort. Other studies have shown decline in retinal thickness over time in aging populations. A study of disease-free controls and MS participants found an average decline of 0.49 mm in pRNFL thickness over 3 years in the healthy control cohort; this is equivalent to a decline of 1.63 mm per decade (28). This study did not account for a faster decline after the age of 40 years, but results are similar to the overall decline found in our study of 1.31 mm per year. A study of normative values in an Asian Indian population showed a similar rate for decline in global pRNFL thickness (1.57 mm per decade) (29). Peripapillary RNFL thinning has been associated with brain atrophy in cognitively normal older adults (30,31) and, thus, may be a normal characteristic of aging. Limitations of this study include that because this is not a population-based study, further evaluation of our conversion equations will be helpful to confirm generalizability. The equations developed in this study are applicable only to the Cirrus OCT or Spectralis device and not interchangeable with other devices. Methods used in the analysis can be applied to develop equations across other OCT devices. Another limitation includes some missing data for race/ ethnicity and for GCIPL thickness. A sensitivity analysis, imputing race/ethnicity and GCIPL thickness, was performed and produced similar results; therefore, it is less likely this missing data created biases in the analysis. Race/ ethnicity was self-reported for this study, which can introduce some error. Furthermore, race and ethnicity are defined differently in the United States compared with other international sites. As such, race/ethnicity data collected from outside of the United States was translated as best as possible to the US categories, which may introduce some measurement error. However, this error is expected to be small and not likely to bias the data, Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 particularly because almost all the non-US participants were classified as Caucasian. Finally, there is limited clinical evaluation in the healthy control participants; these participants have self-reported that they have no known ophthalmological or neurological diseases. No formal ophthalmic assessment or measurement of intraocular pressure (IOP) was performed in most participants. One subset of the healthy control participants at NYU (n = 74) did undergo clinical evaluation. Another subset of participants (n = 139) completed vision-specific quality of life questionnaires (25-Item National Eye Institute Visual Functioning Questionnaire [NEI-VFQ-25]). The scores for these questionnaires were in a range that is consistent with self-reported healthy control status. We, therefore, would not expect a significant amount of bias to be introduced from self-reporting. Ocular axial length has been associated with reduced retinal thickness (32) and with sex (25), but axial length was not measured in this study. However, in a study that was used to develop the Cirrus normative database, refractive error and axial length explained less than 2% of variability for the model (33). Although it is possible that axial length could explain variations in sex differences, not having this measurement in the model is not likely to add significant degrees of bias. Because participants with high-contrast visual acuities worse than 20/40 Snellen equivalent were excluded, and participants self-reported no history of neurological or ophthalmological disease, our control cohort may be healthier than the general population. Results when evaluating a typical aging population, including those with other ocular pathologies including diabetes, high myopia, and macular degeneration, may show varied data for pRNFL and GCIPL thickness. The evolution of SD-OCT presents a unique opportunity to evaluate for optic nerve degeneration in people with MS. Measurements from different devices are not necessarily interchangeable due to differences in segmentation algorithms and acquisition protocols that introduce systematic biases between devices. This may complicate our ability to track disease progression over time if different devices are used. However, a conversion equation or correction factor, as developed in the present investigation, can allow for pooling of data acquired on different devices for research studies. This can also facilitate comparisons of results from different OCT devices. The equations will be helpful, for example, if a patient was scanned clinically on one OCT device and then scanned on a different device subsequently at the next follow-up visit. The availability of normative data by age decade facilitates the evaluation of normal SDOCT thickness values, and age-specific normative data by decade should be considered when assessing for retinal thinning. Rates of decline over time in cohorts of healthy control participants may help to determine if an MS patient, by comparison, has an abnormal rate of retinal thinning. Results from the present study will increase the utility of OCT as a diagnostic tool for optic nerve 451 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution degeneration in MS in particular, and for other neuroophthalmologic disorders in general. 13. STATEMENT OF AUTHORSHIP Conception and design: all authors; Acquisition of data: all authors; Analysis and interpretation of data: L. J. Balcer and R. Kenney. Drafting the manuscript: R. Kenney and L. J. Balcer; Revising the manuscript for intellectual content: all authors. Final approval of the completed manuscript: all authors. 14. 15. REFERENCES 1. Nolan RC, Akhand O, Rizzo JR, Galetta SL, Balcer LJ. Evolution of visual outcomes in clinical trials for multiple sclerosis disease-modifying therapies. J Neuroophthalmol. 2018;38:202–209. 2. Bhargava P, Lang A, Al-Louzi O, Carass A, Prince J, Calabresi PA, Saidha S. Applying an open-source segmentation algorithm to different OCT devices in multiple sclerosis patients and healthy controls: implications for clinical trials. Mult Scler Int. 2015;2015:136295. 3. Pierro L, Gagliardi M, Iuliano L, Ambrosi A, Bandello F. Retinal nerve fiber layer thickness reproducibility using seven different OCT instruments. Invest Ophthalmol Vis Sci. 2012;53:5912– 5920. 4. Miller M, Zemon V, Nolan-Kenney R, Balcer LJ, Goff DC, Worthington M, Hasanaj L, Butler PD. Optical coherence tomography of the retina in schizophrenia: inter-device agreement and relations with perceptual function. Schizophrenia Res. 2020;219:13–18. 5. Watson GM, Keltner JL, Chin EK, Harvey D, Nguyen A, Park SS. Comparison of retinal nerve fiber layer and central macular thickness measurements among five different optical coherence tomography instruments in patients with multiple sclerosis and optic neuritis. J Neuroophthalmol. 2011;31:110–116. 6. Lange AP, Sadjadi R, Costello F, Guber I, Traboulsee AL. Reproducibility of retinal nerve fiber layer measurements with manual and automated centration in healthy subjects using spectralis spectral-domain optical coherence tomography. ISRN Ophthalmol. 2012;2012:860819. 7. Oberwahrenbrock T, Weinhold M, Mikolajczak J, Zimmermann H, Paul F, Beckers I, Brandt AU. Reliability of intra-retinal layer thickness estimates. PLoS One. 2015;10:e0137316. 8. Syc SB, Warner CV, Hiremath GS, Farrell SK, Ratchford JN, Conger A, Frohman T, Cutter G, Balcer LJ, Frohman EM, Calabresi PA. Reproducibility of high-resolution optical coherence tomography in multiple sclerosis. Mult Scler. 2010;16:829–839. 9. Oberwahrenbrock T, Traber GL, Lukas S, Gabilondo I, Nolan R, Songster C, Balk L, Petzold A, Paul F, Villoslada P, Brandt AU, Green AJ, Schippling S. Multicenter reliability of semiautomatic retinal layer segmentation using OCT. Neurol Neuroimmunol Neuroinflamm. 2018;5:e449. 10. Warner CV, Syc SB, Stankiewicz AM, Hiremath G, Farrell SK, Crainiceanu CM, Conger A, Frohman TC, Bisker ER, Balcer LJ, Frohman EM, Calabresi PA, Saidha S. The impact of utilizing different optical coherence tomography devices for clinical purposes and in multiple sclerosis trials. PLoS One. 2011;6:e22947. 11. Buchser NM, Wollstein G, Ishikawa H, Bilonick RA, Ling Y, Folio LS, Kagemann L, Noecker RJ, Albeiruti E, Schuman JS. Comparison of retinal nerve fiber layer thickness measurement bias and imprecision across three spectral-domain optical coherence tomography devices. Invest Ophthalmol Vis Sci. 2012;53:3742–3747. 12. Koh VT, Tham Y-C, Cheung CY, Wong W-L, Baskaran M, Saw SM, Wong TY, Aung T. Determinants of ganglion cell–inner 452 16. 17. 18. 19. 20. 21. plexiform layer thickness measured by high-definition optical coherence tomography. Invest Ophthalmol Vis Sci. 2012;53:5853–5859. Motamedi S, Gawlik K, Ayadi N, Zimmermann HG, Asseyer S, Bereuter C, Mikolajczak J, Paul F, Kadas EM, Brandt AU. Normative data and minimally detectable change for inner retinal layer thicknesses using a semi-automated OCT image segmentation pipeline. Front Neurol. 2019;10:1117. Budenz DL, Anderson DR, Varma R, Schuman J, Cantor L, Savell J, Greenfield DS, Patella VM, Quigley HA, Tielsch J. Determinants of normal retinal nerve fiber layer thickness measured by Stratus OCT. Ophthalmology. 2007;114:1046– 1052. Kimbrough DJ, Sotirchos ES, Wilson JA, Al-Louzi O, Conger A, Conger D, Frohman TC, Saidha S, Green AJ, Frohman EM, Balcer LJ, Calabresi PA. Retinal damage and vision loss in African American multiple sclerosis patients. Ann Neurol. 2015;77:228–236. Caldito NG, Saidha S, Sotirchos ES, Dewey BE, Cowley NJ, Glaister J, Fitzgerald KC, Al-Louzi O, Nguyen J, Rothman A, Ogbuokiri E, Fioravante N, Feldman S, Kwakyi O, Risher H, Kimbrough D, Frohman TC, Frohman E, Balcer L, Crainiceanu C, Van Zijl PCM, Mowry EM, Reich DS, Oh J, Pham DL, Prince J, Calabresi PA. Brain and retinal atrophy in African-Americans versus Caucasian-Americans with multiple sclerosis: a longitudinal study. Brain. 2018;141:3115–3129. Balcer LJ, Balk LJ, Brandt AU, Calabresi PA, Martinez-Lapiscina EH, Nolan RC, Paul F, Petzold A, Saidha S. The international multiple sclerosis visual system consortium: advancing visual system research in multiple sclerosis. J Neuroophthalmol. 2018;38:494–501. Martinez-Lapiscina EH, Arnow S, Wilson JA, Saidha S, Preiningerova JL, Oberwahrenbrock T, Brandt AU, Pablo LE, Guerrieri S, Gonzalez I, Outteryck O, Mueller AK, Albrecht P, Chan W, Lukas S, Balk LJ, Fraser C, Frederiksen JL, Resto J, Frohman T, Cordano C, Zubizarreta I, Andorra M, SanchezDalmau B, Saiz A, Bermel R, Klistorner A, Petzold A, Schippling S, Costello F, Aktas O, Vermersch P, Oreja-Guevara C, Comi G, Leocani L, Garcia-Martin E, Paul F, Havrdova E, Frohman E, Balcer LJ, Green AJ, Calabresi PA, Villoslada P; IMSVISUAL Consortium. Retinal thickness measured with optical coherence tomography and risk of disability worsening in multiple sclerosis: a cohort study. Lancet Neurol. 2016;15:574–584. Cruz-Herranz A, Balk LJ, Oberwahrenbrock T, Saidha S, Martinez-Lapiscina EH, Lagreze WA, Schuman JS, Villoslada P, Calabresi P, Balcer L, Petzold A, Green AJ, Paul F, Brandt AU, Albrecht P; IMSVISUAL Consortium. The APOSTEL recommendations for reporting quantitative optical coherence tomography studies. Neurology. 2016;86:2303–2309. Aytulun A, Cruz-Herranz A, Aktas O, Balcer LJ, Balk L, Barboni P, Blanco AA, Calabresi PA, Costello F, Sanchez-Dalmau B, DeBuc DC, Feltgen N, Finger RP, Frederiksen JL, Frohman E, Frohman T, Garway-Heath D, Gabilondo I, Graves JS, Green AJ, Hartung HP, Havla J, Holz FG, Imitola J, Kenney R, Klistorner A, Knier B, Korn T, Kolbe S, Kramer J, Lagreze WA, Leocani L, Maier O, Martinez-Lapiscina EH, Meuth S, Outteryck O, Paul F, Petzold A, Pihl-Jensen G, Preiningerova JL, Rebolleda G, Ringelstein M, Saidha S, Schippling S, Schuman JS, Sergott RC, Toosy A, Villoslada P, Wolf S, Yeh EA, Yu-Wai-Man P, Zimmermann HG, Brandt AU, Albrecht P. APOSTEL 2.0 recommendations for reporting quantitative optical coherence tomography studies. Neurology. 2021;97:68–79. Nolan-Kenney RC, Liu M, Akhand O, Calabresi PA, Paul F, Petzold A, Balk L, Brandt AU, Martinez-Lapiscina EH, Saidha S, Villoslada P, Al-Hassan AA, Behbehani R, Frohman EM, Frohman T, Havla J, Hemmer B, Jiang H, Knier B, Korn T, Leocani L, Papadopoulou A, Pisa M, Zimmermann H, Galetta SL, Balcer LJ; International Multiple Sclerosis Visual System Consortium. Optimal intereye difference thresholds by optical coherence tomography in multiple sclerosis: an international study. Ann Neurol. 2019;85:618–629. Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution 22. Tewarie P, Balk L, Costello F, Green A, Martin R, Schippling S, Petzold A. The OSCAR-IB consensus criteria for retinal OCT quality assessment. PLoS One. 2012;7:e34823. 23. Bollen KA. Structural Equations with Latent Variables. New York: Wiley, 1989:514. xiv. 24. Li D, Rauscher FG, Choi EY, Wang M, Baniasadi N, Wirkner K, Kirsten T, Thiery J, Engel C, Loeffler M, Elze T. Sex-specific differences in circumpapillary retinal nerve fiber layer thickness. Ophthalmology. 2020;127:357–368. 25. Mwanza JC, Durbin MK, Budenz DL, Girkin CA, Leung CK, Liebmann JM, Peace JH, Werner JS, Wollstein G; Cirrus OCT Normative Database Study Group. Profile and predictors of normal ganglion cell-inner plexiform layer thickness measured with frequency-domain optical coherence tomography. Invest Ophthalmol Vis Sci. 2011;52:7872–7879. 26. Poon LYC, Antar H, Tsikata E, Guo R, Papadogeorgou G, Freeman M, Khoueir Z, Lee R, Shieh E, Simavli H, Que CJ, de Boer JF, Chen TC. Effects of age, race, and ethnicity on the optic nerve and peripapillary region using spectral-domain OCT 3D volume scans. Transl Vis Sci Technol. 2018;7:12. 27. Alasil T, Wang K, Keane PA, Lee H, Baniasadi N, de Boer JF, Chen TC. Analysis of normal retinal nerve fiber layer thickness by age, sex, and race using spectral domain optical coherence tomography. J Glaucoma. 2013;22:532–541. Kenney et al: J Neuro-Ophthalmol 2022; 42: 442-453 28. Talman LS, Bisker ER, Sackel DJ, Long DA Jr, Galetta KM, Ratchford JN, Lile DJ, Farrell SK, Loguidice MJ, Remington G, Conger A, Frohman TC, Jacobs DA, Markowitz CE, Cutter GR, Ying GS, Dai Y, Maguire MG, Galetta SL, Frohman EM, Calabresi PA, Balcer LJ. Longitudinal study of vision and retinal nerve fiber layer thickness in multiple sclerosis. Ann Neurol. 2010;67:749–760. 29. Appukuttan B, Giridhar A, Gopalakrishnan M, Sivaprasad S. Normative spectral domain optical coherence tomography data on macular and retinal nerve fiber layer thickness in Indians. Indian J Ophthalmol. 2014;62:316–321. 30. Shi Z, Cao X, Hu J, Jiang L, Mei X, Zheng H, Chen Y, Wang M, Cao J, Li W, Li T, Li C, Shen Y. Retinal nerve fiber layer thickness is associated with hippocampus and lingual gyrus volumes in nondemented older adults. Prog Neuropsychopharmacol Biol Psychiatry. 2020;99:109824. 31. Shi Z, Zheng H, Hu J, Jiang L, Cao X, Chen Y, Mei X, Li C, Shen Y. Retinal nerve fiber layer thinning is associated with brain atrophy: a longitudinal study in nondemented older adults. Front Aging Neurosci. 2019;11:69. 32. Jonas JB, Xu L, Wei WB, Pan Z, Yang H, Holbach L, PandaJonas S, Wang YX. Retinal thickness and axial length. Invest Ophthalmol Vis Sci. 2016;57:1791–1797. 33. Carl Zeiss Meditec 2015; 501K Summary. 453 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.